MICCAI 24
Incredible Experience in Marrakesh. We presented six works on anomaly detection and won 3 awards!
NEWS
10/10/20241 min read
As the 27th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2024) concludes, I’m grateful for the opportunity to have been part of such a stimulating event. It was a privilege to present our work on the reverse process of diffusion models for anomaly detection, focusing on THOR (Temporal Harmonization for Optimal Restoration) during the main conference.
I'm also deeply appreciative of the recognition our group's efforts received, including:
2nd Place in the PhD Thesis Madness reflected our contributions to unsupervised anomaly detection.
Best Paper Award for our work on Deformable Autoencoders for Alzheimer's Disease Analysis, offering a new approach to localizing and evaluating anomalies in brain MRI.
Best Paper Award for Selective Test-Time Adaptation for Anomaly Detection, where we explored how adaptive models could improve anomaly detection across diverse data.
Our group also presented other notable contributions, including
The combination of Language Models with Anomaly Detection enhances interpretability and generalizability.
The use of Diffusion Models for Fetal Brain MRI opens up new perspectives for anomaly detection in early-stage diagnostics.
Counterfactual Pathology Synthesis, exploring how realistic synthetic pathology data can be generated.
MICCAI 2024 provided a fantastic platform for exchanging ideas, and I look forward to seeing how these developments will advance the field of medical imaging.